Intra Prediction Filtering for Lower-Complexity Video Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards face challenges in achieving a balance between coding and decoding performance and time complexity due to inefficiencies in intra-prediction techniques based on interpolation.
Innovation Solution
A method for determining a target filtering mode, reference region, and filtering coefficients for intra prediction, which adapts to the size of the current block, improving encoding and decoding efficiency by optimizing the reference region used for calculating filtering coefficients.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If interpolation filtering coefficients are obtained using reconstructed sample values around a current block for intra-prediction, then intra-prediction accuracy is improved, but time complexity increases
Solution Approach 1:
The patent applies local quality by selecting different reference regions and filtering modes based on the specific characteristics of the current block. Instead of using a uniform approach for all blocks, the method adapts the reference region selection and filtering coefficients to local block properties, achieving high prediction accuracy for each specific block while optimizing computational resources.
Solution Approach 2:
The patent implements dynamics by making the reference region selection and filtering mode adaptive rather than static. The system dynamically determines the appropriate reference region and filtering coefficients based on the current block's characteristics, allowing the intra-prediction process to adapt to varying block conditions and optimize both accuracy and efficiency.
2Device complexity
If a fixed reference region is used for calculating filtering coefficients, then device complexity is reduced, but intra-prediction accuracy deteriorates
Solution Approach 1:
The patent transforms the static reference region approach into a dynamic one by introducing adaptive selection mechanisms. The reference region is determined based on the current block's characteristics, enabling the system to optimize prediction accuracy for different block types while maintaining manageable device complexity through algorithmic adaptation rather than hardware complexity.
Solution Approach 2:
The patent applies parameter changes by varying the reference region parameters and filtering coefficients according to the current block's properties. Instead of using fixed parameters for all blocks, the system adjusts these parameters dynamically, achieving high prediction accuracy across different block conditions without requiring complex device architecture.
3Measurement precision
If adaptive reference region selection is implemented based on current block size, then intra-prediction accuracy is improved, but device complexity increases
Solution Approach 1:
The patent applies local quality by tailoring the reference region selection to the specific size and characteristics of each current block. This localized adaptation ensures that each block receives optimized prediction processing appropriate to its dimensions, improving overall prediction accuracy while keeping the complexity management localized rather than requiring system-wide complexity increases.
Solution Approach 2:
The patent implements parameter changes by adjusting reference region parameters based on the current block size. This approach allows the system to optimize prediction accuracy for different block dimensions through parameter adaptation rather than through complex device architecture, achieving high performance with controlled complexity.
Data Source
AI summary
Embodiments of the present application disclose an encoding method, a decoding method, and a storage medium. The decoding method, applied to a decoder, includes: decoding a bitstream, and determining a target filtering mode for a current block; determining a reference area for the current block on the basis of a size parameter of the current block and the target filtering mode; determining filter coefficients for the current block on the basis of the reference area for the current block; and performing intra prediction on the current block on the basis of the filter coefficients, and determining predicted values of the current block.